Mycobacterium porcinum causing panniculitis in the cat

Abstract

A 4-year-old domestic shorthair cat was presented with a 7-month history of nodules and draining fistulous tracts of the ventral abdomen. Histopathological examination of affected tissue revealed acid-fast bacilli stained by the Ziehl-Neelsen procedure. Deep tissue culture confirmed infection with a rapidly growing mycobacterium, and gene sequencing characterized the organism as Mycobacterium porcinum. Treatment with pradofloxacin and doxycycline resulted in clinical resolution of the lesions. On continued antibiotic therapy 7 months later, there was no local recurrence nor were there clinical signs associated with distant spread of the infection. This is the first clinical description of a feline infection with this organism.

Key clinical message:

This is the first clinical description of mycobacterial panniculitis in a cat due to genetically characterized Mycobacterium porcinum. This case report highlights a disease entity that can present a diagnostic and therapeutic challenge to clinicians.

Mycobacteria are widespread, aerobic, non-spore-forming, non-motile bacilli that are implicated as a cause of various clinical syndromes in cats, dogs, cattle, and humans. This includes localized cutaneous disease, pyogranulomatous lobar pneumonia, and very rarely, disseminated systemic infections (1–3). Of these, localized infections of the skin and subcutis (mycobacterial panniculitis) caused by rapidly growing mycobacteria is the most common presentation (1,2). Rapidly growing mycobacteria are so called due to their ability to grow on appropriate media within 7 d of incubation. These organisms, also classified as atypical or non-tuberculous mycobacteria, display a predilection for adipose tissue sites due to the high triglyceride content which provides energy for growth (1). Feline mycobacterial panniculitis most typically affects the inguinal and ventral abdominal fat pads following traumatic skin damage and subsequent wound inoculation, and tends to occur in immunocompetent individuals (1,2,4,5). Integumentary breaches such as bite wounds, penetrating foreign bodies, injections, or surgery can allow these bacteria to overcome host defenses (1,2,4). Lesions can develop over a prolonged clinical course and can initially appear as circumscribed nodules with focal thickening of the subcutis, progressing to the development of punctate fistulae (2). Atypical mycobacteriosis is considered non-zoonotic, though rare transmission of rapidly growing mycobacteria has been documented (6). Defining the mycobacterial species involved with culture and molecular techniques is necessary to distinguish from potentially zoonotic mycobacterial species (7). The rapidly growing mycobacteria are ubiquitous in nature, especially in water and wet soil, and are non-pathogenic for animals under normal circumstances (1,5). The disease has a low incidence and it can present diagnostic and therapeutic challenges due to the prolonged clinical course, lack of pathognomonic histopathological changes, and failure to culture many mycobacteria. The case report here highlights the disease process and is the first characterization of a feline infection genetically confirmed to be due to Mycobacterium porcinum.

Case description

A 4-year-old, 5.6 kg, spayed, female domestic shorthair cat was referred for investigation of draining nodules affecting the skin of the ventral abdomen. Seven months previously, the cat had been presented to the referring veterinarian for evaluation of an abdominal wound. The cat had been rescued from Portugal 1 mo before the development of skin lesions, and otherwise appeared to be in good health, with no history of trauma or recurrent infections. There was also no history of illness of the owners or other pets in the same household. Following initial assessment of the abdominal wound by the referring veterinarian, the cat was treated with cefovecin (Convenia; Zoetis Canada, Kirkland, Quebec), 8 mg/kg body weight (BW), SC, once, and meloxicam (Metacam; Boehringer Ingelheim Canada, Burlington, Ontario), 0.01 mg/kg BW, SC, then 0.05 mg/kg BW, PO, q24h for 5 d. In the 2 wk following this treatment, the cat developed multiple draining fistulous tracts of the abdominal fat pad. Treatment was then initiated with amoxicillin clavulanic acid (Clavaseptin; Vétoquinol Canada, Lavaltrie, Quebec), 9.7 mg/kg BW, PO, q12h for 14 d, along with cleansing of the affected areas using a 1% chlorhexidine solution q12h. Two weeks later, following further progression of the draining lesions on the abdomen, surgical debridement and primary closure (PDS; Ethicon, Johnson and Johnson, Bridgewater, New Jersey, USA) with placement of Penrose drains was performed in the affected area (8). Post-surgical treatments included metronidazole (Summit Veterinary Pharmacy, Aurora, Ontario), 12.5 mg/kg BW, PO, q12h for 6 d and enrofloxacin (Baytril; Bayer Canada, Mississauga, Ontario), 2.5 mg/kg BW, PO, q24h for a total of 17 d. Two months after surgery, a relapse of the lesions occurred. Based on a differential diagnosis of eosinophilic granuloma complex, prednisolone (prednisolone; Rafter 8 Products, Calgary, Alberta) was started at 0.44 mg/kg BW, PO, q12h for 7 d, then 0.44 mg/kg BW q24h for 7 d; this was subsequently increased to 0.86 mg/kg BW q12h for relief of pruritus, and this treatment was continued until referral.

One month before referral, blood tests including a complete blood (cell) count (CBC) and serum biochemistry profile were performed. The hemogram detected mild reticulocytosis [66.4 × 10³/μL; reference range (RR): 3 to 50 × 10³/μL] with an erythrocyte count within the reference range (red blood cell count 8.3 × 10¹²/L; RR: 7.1 to 11.5 × 10¹²/L). The leukocyte count was within the reference range (5.6 × 10⁹/L; RR: 3.9 to 19.0 × 10⁹/L) and the platelet count was 289 × 10⁹/L; RR: 155 to 641 × 10⁹/L. Biochemical blood tests showed an elevated blood glucose (12.1 mmol/L; RR: 4.0 to 9.7 mmol/L) suspected to be consistent with a stress response, mild hypochloremia (108 mmol/L; RR: 114 to 126 mmol/L), and mildly elevated alkaline phosphatase (92 IU/L; RR: 12 to 59 IU/L) with a total bilirubin within the normal reference range (1.6 μmol/L; RR: 0 to 5.13 μmol/L). A tissue biopsy was also collected from the affected region at the time of blood testing. Dermatohistopathology of the tissue samples revealed a pyogranulomatous dermatitis and panniculitis with fibrosis and intralesional rod-shaped bacteria (Figure 1A). Acid-fast (Ziehl-Neelsen) staining revealed moderate numbers of acid-fast bacilli both extracellularly (Figure 1B) and intracellularly within macrophages. Periodic Acid Schiff (PAS) staining was negative for fungal elements. Prior to referral, amoxicillin clavulanic acid (Clavaseptin; Vétoquinol Canada), 13 mg/kg BW, PO, q12h, for 14 d, and enrofloxacin (Baytril; Bayer Canada), 5.2 mg/kg BW, PO, q12h for 7 d, were also added to the treatment.

Figure 1.

A — Histology of pyogranulomatous inflammation (Hematoxylin-eosin stain; original magnification ×100) encapsulating mycobacterial rods (arrow) within a lipocyst. B — Ziehl-Neelsen staining of acid-fast bacilli (Mycobacterium porcinum) (arrow) within the panniculus surrounded by neutrophils and macrophages.

Seven months after the initial presentation, the cat was presented for a dermatologic opinion. On presentation, the cat was found to be afebrile (37.4°C; RR: 36.7°C to 38.9°C) and in good body condition. No systemic clinical signs were reported by the owner. Examination of the skin revealed firm, fibrotic nodules on the lateral sides of the abdominal fat pad extending dorsally to the level of the lateral abdomen. There was palpable edema of the right lateral abdomen, and the ventral abdomen displayed a few foci of ulceration consistent with early fistulous tract formation. Cytologic examination of impressions from the ventral abdomen, stained with a commercial modified Wright’s stain (DipQuick; Jorgensen Laboratories, Loveland, Colorado, USA) were assessed on 100× oil immersion power and revealed neutrophils and epithelioid histiocytes. An in-clinic SNAP enzyme-linked immunosorbent assay (ELISA) for feline leukemia virus and feline immunodeficiency virus (IDEXX Laboratories, Westbrook, Maine, USA) was negative.

Based on the finding of acid-fast bacilli and pyogranulomatous inflammation on previous dermatohistopathology, deep tissue biopsies were collected for a macerated tissue culture to attain a definitive diagnosis of mycobacterial panniculitis and target antibiotic treatment. For sample collection, the cat was sedated with midazolam (Midazolam; Sandoz, Boucherville, Quebec), 0.2 mg/kg BW, IV, and butorphanol (Torbugesic; Zoetis Canada), 0.2 mg/kg BW, IV, and anesthetized with propofol (Propoflo 28; Zoetis Canada), 3.2 mg/kg BW, IV. Pending the results of tissue culture and the suspected mycobacterial etiology, the cat was started on antibiotic treatments of marbofloxacin (Zeniquin; Zoetis Canada), 4.4 mg/kg BW, PO, q24h, and clarithromycin (Apo-Clarithromycin; Apotex, Toronto, Ontario), 11 mg/kg BW, PO, q12h. Prednisolone was gradually tapered with a 50% dose reduction every 2 wk until it was stopped.

The skin samples were sent to a local reference laboratory for aerobic, anaerobic, and fungal cultures, and to a laboratory specializing in the isolation of mycobacteria for mycobacterial cultures. Aerobic cultures were positive for Corynebacterium species, which was susceptible to both clarithromycin and doxycycline. Corynebacteria are Gram-positive bacilli that can be cultured as normal flora from the cat (9). Anaerobic and fungal cultures were negative for growth.

The skin lesions continued to progress in the 4 wk following initial presentation, leading to the development of approximately 10 new draining tracts on the ventral abdomen (Figure 2A). As mycobacterial culture results were still pending, an empiric change of antibiotics to pradofloxacin (Veraflox; Bayer Canada), 7.7 mg/kg BW, PO, q24h, and doxycycline (Summit Veterinary Pharmacy), 9.7 mg/kg BW, PO, q12h was started based on the poor response to initial treatments.

Figure 2.

A — The ventral abdomen of a 4.5-year-old female spayed domestic shorthair cat following a relapse of draining lesions representative of mycobacterial panniculitis. B — The ventral abdomen of the cat at a 7-month post-treatment follow-up.

To culture for mycobacteria, the skin samples were pretreated with 1.5% KOH as a decontaminant before inoculating separately onto modified Middlebrook 7H11 agar (with pyruvate and glycerol), Stonebrink’s medium, Löwenstein–Jensen medium, and mycobacterium growth indicator tube (MGIT; Becton Dickinson, Baltimore, Maryland, USA). After 5 d of incubation at 37°C, substantial growth was observed on all media. Antimicrobial susceptibility testing as recommended for nontuberculous mycobacteria (10) indicated susceptibility to moxifloxacin (a third-generation fluoroquinolone) but resistance to doxycycline at an MIC of > 6 μg/mL. For species identification, DNA from the isolate was extracted using Quick-DNA kit (Zymo Research, Irvine, California, USA). The isolate was then sequenced at the 16S rRNA and the rpoB regions of the genome, as it has been demonstrated that rpoB gene sequencing provides higher discriminatory power to distinguish nontuberculous mycobacteria species and subspecies (11). The gene sequences were deposited in GenBank and compared with known 16S rRNA and rpoB sequences using BLAST. The 16S rRNA (825 bp) sequence had a 100% match with M. porcinum and M. neworleansense (both members of the Mycobacterium fortuitum third biovariant complex). The sequence of the rpoB gene (693 bp) showed a 99% match with M. porcinum. No other identification methods were applied. Culture isolation and gene sequencing was performed by the National Veterinary Services Laboratories, United States Agriculture Department in Ames, Iowa, USA.

In the ensuing weeks, pending the final mycobacterial culture results, the cat’s draining tracts had resolved, and no new lesions were forming with the prescribed combination of pradofloxacin and doxycycline. Additional susceptibility testing with pradofloxacin could not be performed by the testing laboratory at this time, but the isolate was susceptible in vitro to moxifloxacin (a human third-generation fluoroquinolone equivalent), and the patient was responding well to pradofloxacin. Pradofloxacin has been demonstrated to be effective for treatment of rapidly growing mycobacteria (12). The isolate was resistant in vitro to doxycycline but the decision was made to continue the doxycycline based on the patient’s positive progress and in an effort to prevent relapse by continuing all treatments that had resulted in clinical improvement. Doxycycline was selected initially as it is commonly recommended to treat rapidly growing mycobacteria with a combination of one or more oral antibiotics due to potential for antimicrobial resistance to develop with fluoroquinolone monotherapy (1,12,13). Interestingly, the susceptibility results indicated susceptibility to the previously used clarithromycin, and an intermediate susceptibility to ciprofloxacin (spectrum of activity similar to marbofloxacin). The reason for the patient’s lack of response to the previous combination of antibiotics is unknown, although immunosuppression from the initial empiric use of prednisolone may have played a role in the poor response to marbofloxacin and clarithromycin. This highlights the importance of ruling out infection before initiating empiric treatment with prednisolone, as glucocorticoids are contraindicated in the treatment of mycobacterial panniculitis. In general, very good agreement has been observed in previous studies between in vitro susceptibility testing and in vivo response to antimicrobials in the treatment of mycobacterial panniculitis in cats (14).

At both a 5-month and a 7-month follow-up, the cat had residual alopecia, scarring, and palpable fibrotic bands within the subcutis, but no appreciable nodules or draining tracts had reformed (Figure 2B). Treatment of mycobacterial panniculitis is recommended for a minimum of 3 to 12 mo, with therapy continuing at least 1 to 2 mo after affected tissues look and feel completely normal (1,15). As such, treatment recommendations were to continue daily doxycycline and pradofloxacin therapy as prescribed until another recheck scheduled 4 mo later.

Discussion

Mycobacterial infections remain an uncommon cause of cutaneous infection in cats, however, international prevalence is variable (4,13,14). Cats appear to be at greater risk of developing infection with rapidly growing mycobacteria than other domestic species, especially if involved in hunting or fighting, which can increase opportunities for wound inoculation (4,14). The opportunistic rapidly growing mycobacteria more commonly identified in feline pyogranulomatous panniculitis include Mycobacterium smegmatis, M. fortuitum, M. chelonae, and M. phlei (1,5,15). Infections involving other rapidly growing mycobacteria species have also been documented in cats, including M. abscessus, M. flavescens, M. thermoresistibile, M. goodii, M. mageritense, and M. alvei (12,16–20).

Mycobacterium porcinum is a rarely encountered RGM, first isolated in 1973 from porcine lymph nodes showing tuberculosis-like lymphadenitis (21). It was further characterized as being similar to M. fortuitum in 1983 and deposited in the American Type Culture collection (ATCC) as type strain ATCC 33776 (22). Mycobacterium porcinum has since been further classified as a member of the third biovariant group of the M. fortuitum complex, including 4 other isolates: M. boenickei, M. houstonense, M. neworleansense, and M. brisbanense (23). Mycobacterium porcinum infections in humans have been isolated from wounds, central catheter infections (with or without bacteremia), in pneumonitis, and in peritonitis associated with peritoneal dialysis (23–26). M. porcinum has also been isolated from bovine bulk milk supply, and infections have been linked to contaminated community and hospital water sources (24,27). Mycobacterium porcinum has been previously isolated from 2 cats in Australia, though not clinically characterized in the affected individuals (12).

Atypical mycobacterial infections in cats can be seen at any age (range: 2 to 12 y) and there appears to be a predilection for overweight females to be affected (4,28). As mycobacteria display a tropism for lipid-rich tissue, localization to the inguinal fat pad is observed in the preponderance of affected cats, as in this case (4). The distinguishing characteristic of all Mycobacterium species are the thick, hydrophobic cell walls which are rich in mycolic acids and render the organisms acid-fast when stained with Ziehl-Neelsen or Fite-Faraco stain (1). This cell wall provides a substantial contribution to the hardiness of the organism, allowing survival within phagocytes, and resulting in the typical granulomatous immune response by hosts (1). When presented with a cat displaying 1 or more non-healing abscesses localizing to the abdomen, especially those failing to respond to standard antimicrobial therapy and wound management, mycobacterial panniculitis should be considered as a top differential diagnosis. If suspected, fine-needle aspirates for cytology and biopsies for histopathology (including acid-fast staining), with macerated tissue cultures for aerobic, anaerobic, fungal, and mycobacterial organisms should be pursued to obtain a definitive diagnosis and to rule out other subcutaneous infections, such as deep mycosis. A more rapid identification technique includes fine-needle aspiration of subcutaneous nodules through previously disinfected intact skin, followed by inoculation of material directly into mycobacterial culture media or blood culture medium within a few minutes of sample collection (1). Cytological preparations of fine-needle samples can also rapidly identify organisms with acid fast staining, or as “negative images” on modified Wright-stained (Diff-Quik) slides (29). Owners should be counselled that some mycobacteria may fail to be cultured, and this does not preclude mycobacterial infection (13,28). Additionally, acid-fast rods may be difficult to identify on histopathology, and have not been identified in some samples that later cultured positive for mycobacteria (4,28). Reference laboratories with experience in the culture and detection of mycobacteria should be selected (1,28).

Species identification and susceptibility data for rapidly growing mycobacteria can significantly impact antimicrobial selection, and antibiotic resistance in rapidly growing mycobacteria is increasingly common (1,5). When a tentative diagnosis of mycobacterial panniculitis is made, it is advantageous to start empirical antimicrobial treatment with one or more antibiotics considered widely effective for rapidly growing mycobacteria prior to susceptibility results to avoid a progression of the condition (1). This involves consideration of the general susceptibility patterns established by retrospective data for RGM. In general, members of the M. fortuitum group are identified more commonly in North America but have a worldwide distribution; previous characterizations of M. fortuitum infections indicate they are often susceptible to clarithromycin and pradofloxacin (1,12,14,20). Treating at the high end of the antibiotic dosing range is recommended as subcutaneous tissues are not well-perfused, and this may impact antimicrobial diffusion from the blood to organisms in the panniculus (1). Owners should be prepared for a regimen of 1 or more antibiotics administered over several months in the management of mycobacterial panniculitis (1,14,30). Some experienced clinicians recommend en bloc surgical resection of affected tissue in correlation with antimicrobial therapy, given that adequate levels of the antimicrobial agent may not be reached in the involved tissues (1,14). However, animals treated in a preliminary fashion using oral antibiotics can progressively respond to the extent that surgery can be avoided, or can then be focused on more localized recalcitrant lesions if medical therapy has plateaued (1). Surgery without appropriate antimicrobial therapy also risks the spread of lesions and subsequent relapse, as in this case.

In summary, some patients may require long-term antibiotic therapy to maintain remission, and relapse may be seen following discontinuation of treatment (1,5). The prognosis for mycobacterial panniculitis with rapidly growing mycobacteria is considered guarded; however, this condition is becoming increasingly treatable as molecular techniques enhance identification of organisms, and our ability to recognize, diagnose, and manage this condition continues to evolve and improve (1). This paper highlights a patient with a good response to treatment following a definitive diagnosis of the infection, and provides clinical characterization of a genetically confirmed M. porcinum infection in a cat.